ASSISTING PARALLELIZATION OF A COMPUTER PROGRAM

• US 20150089468A1
• Filed: 09/20/2013
• Published: 03/26/2015
• Est. Priority Date: 09/20/2013
• Status: Active Grant

• Current Assignee: CRAY Incorporated (Hewlett-Packard Enterprise Company)
• Original Assignee: CRAY Incorporated (Hewlett-Packard Enterprise Company)
• Inventors: DeRose, Luiz, Poxon, Heidi, Levesque, John, Johnson, Brian H.
• Granted Patent: US 9,250,877 B2
• Time in Patent Office: Days
• US Class Current: 717/106
• CPC Class Codes: G06F 11/3404 for parallel or distributed programming
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 8/30 Creation or generation of source code
• CPC Class Codes: G06F 8/314 Parallel programming languages G06F8/313 takes precedence
• CPC Class Codes: G06F 8/423 Preprocessors
• CPC Class Codes: G06F 8/443 Optimisation
• CPC Class Codes: G06F 8/4443 Inlining
• CPC Class Codes: G06F 8/456 Parallelism detection

Abstract

A parallelization assistant tool system to assist in parallelization of a computer program is disclosed. The system directs the execution of instrumented code of the computer program to collect performance statistics information relating to execution of loops within the computer program. The system provides a user interface for presenting to a programmer the performance statistics information collected for a loop within the computer program so that the programmer can prioritize efforts to parallelize the computer program. The system generates inlined source code of a loop by aggressively inlining functions substantially without regard to compilation performance, execution performance, or both. The system analyzes the inlined source code to determine the data-sharing attributes of the variables of the loop. The system may generate compiler directives to specify the data-sharing attributes of the variables.

Claims

1. A method performed by a computing device for assisting in parallelization of a computer program, the computer program having source code, the method comprising:

executing instrumented code of the computer program to collect performance statistics information relating to execution of loops within the computer program;

presenting to a user performance statistics information collected for a loop within the computer program;

generating inlined source code of the loop, the inlined source code generated by aggressively inlining invoked functions substantially without regard to compilation performance, execution performance, or both;

analyzing the inlined source code to determine the data-sharing attributes of the variables of the loop;

providing feedback on issues found when determining data-sharing attributes of the variables of the loop; and

generating compiler directives to specify the data-sharing attributes of the variables.

2. The method of claim 1 wherein the performance statistics information for a loop includes an execution time derived from a start time indicating time at which execution of the loop was started and an end time indicating time at which execution of the loop ended.

3. The method of claim 2 wherein the execution time is derived from the start time and end time of each execution of the loop.

4. The method of claim 1 wherein the performance statistics information includes, for each execution of the loop, number of iterations of that execution of the loop.

5. The method of claim 1 wherein the performance statistics information includes number of times the loop was executed.

6. The method of claim 1 wherein the computer program is instrumented during compilation by a compiler inserting code to collect the performance statistics information.

7. The method of claim 6 wherein the source code is not instrumented to collect the statistical information.

8. The method of claim 1 including inserting the compiler directives into the source code.

9. The method of claim 1 wherein the analyzing determines that the data-sharing attribute of a variable cannot be resolved.

10. The method of claim 9 including generating an invalid compiler directive indicating that a variable is unresolved, the invalid compiler directive preventing compilation of the source code by a compiler and serving as an indication of variables that are unresolved during editing of the source code.

11. The method of claim 9 including presenting to the user an indication of a reason why a variable cannot be resolved.

12. A computer-readable storage medium storing computer-executable instructions for controlling a computing device to determine a data-sharing attribute of variables of a computer program, the computer-executable instructions comprising instructions that:

generate inlined source code for a loop, the inlined source code generated by aggressively inlining invoked functions substantially without regard to compilation performance or execution performance;

analyze the inlined source code to resolve the data-sharing attributes of the loop, the analysis indicating whether a data-sharing attribute of a variable is shared, private, or unresolved; and

output an indication of whether the data-sharing attributes of the variables are shared, private, or unresolved.

13. The computer-readable storage medium of claim 12 including instructions that generate compiler directives to specify the data-sharing attributes of the variables.

14. The computer-readable storage medium of claim 13 which the instructions that generate a compiler directive generate an invalid compiler directive to indicate that the data-sharing attribute of a variable is unresolved, the invalid compiler directive to prevent compilation of the source code by a compiler and to serve during editing of the source code as an indication that the data-sharing attribute of a variable is unresolved.

15. The computer-readable storage medium of claim 12 including instructions that present to a user an indication of a reason why the data-sharing attribute of a variable is unresolved.

16. A computing system for assisting in parallelization of a computer program, the computer program having source code, the computing system comprising:

one or more computer-readable storage media storing computer-executable instructions that

present to a user performance statistics information collected for loops of the computer program, the performance statistics information for allowing the user to prioritize efforts to parallelize the computer program;

generate an inlined source code of the loop, the inlined source code generated by aggressively inlining invoked functions substantially without regard to compilation performance, execution performance, or both;

analyze the inlined source code to resolve data-sharing attributes of variables of the loop; and

output an indication of whether the data-sharing attributes of variables are shared, private, or unresolved; and

a processor that executes the computer-executable instructions stored in the one or more computer-readable storage media.

17. The computing system of claim 16 including instructions that generate compiler directives to specify the data-sharing attributes of variables as shared or private.

18. The computing system of claim 16 including instructions that generate an invalid compiler directive indicating that the data-sharing attribute of a variable is unresolved, the invalid compiler directive preventing compilation of the source code by a compiler and serving during editing of the source code as an indication of variables whose data-sharing attributes are unresolved.

19. The computing system of claim 16 including instructions that direct execution of an instrumented code of the computer program to collect performance statistics information relating to execution of loops within the computer program.

20. The computing system of claim 19 wherein the source code of the computer program is not instrumented to collect the performance statistics information and wherein the computer program is instrumented by a compiler that inserts code into the compiled code for collecting the performance statistics information.